Mechanism of Atmospheric Oxidation of Sulfur Dioxide by Hydroxyl

Chapter 26

Mechanism of Atmospheric Oxidation of Sulfur Dioxide by Hydroxyl Radicals

Downloaded by UNIV OF IOWA on September 2, 2016 | http://pubs.acs.org Publication Date: April 27, 1989 | doi: 10.1021/bk-1989-0393.ch026

Larry G. Anderson, Paul M . Gates, and Charles R. Nold Department of Chemistry, University of Colorado-Denver, Denver, CO 80204 The H O S O radical formed in the reaction of hydroxyl radicals with SO was found to react with both O and NO. The kinetics of these two reactions were studied in a discharge flow-resonance fluorescence system. The rate constant for the H O S O + O reaction was found to be about (4 -2) x 10 cm /s, in good agreement with previously reported values. The rate constant for the HOSO + NO reaction was found to be about (1 -.5)x 10 cm /s, which is in between the two values previously reported for this rate constant. The H O S O + NO reaction is too slow to be of significance in the atmosphere, when compared to the HOSO + O reaction. This latter reaction forms a HO radical, hence the H O reaction with S O does not act as a radical sink under atmospheric conditions. 2

2

2

2

+4

-13

2

3

+1

-12

2

3

2

2

2

2

2

The oxidation of sulfur dioxide is of importance not only in the polluted atmosphere, but it is also of importance in the natural atmosphere. Toon et al. (1) have recently published a paper on the sulfur cycle in the marine atmosphere. In that work, they conclude that there are no known natural emissions of SO2 to the marine atmosphere. Yet, based on the measured concentrations of SOo and the expected lifetime of SO2 in the marine atmosphere, there must oe a source of SO3 of about 30 Tg S y r in the marine environment (1). The vast majority of this SO^ is believed to arise from the atmospheric oxidation of dimethyl sulfide, which has an estimated biogenic source strength of 40 ± 20 Tg S y r ' in the marine environment (I). The atmospheric oxidation of other biogenic sulfur containing compounds are also believed to contribute to the SO2 production. These compounds include carbonyl sulfide, carbon disulfide and hydrogen sulfide. But these compounds are not expected to contribute significantly to the production of SO2 in the marine environment, since the combined biogenic source strengths for these compounds is believed to be only a few Tg S y r . This budget information suggests that most of the biogenic sulfur emissions in the marine atmosphere are oxidized to SO2. Hence, it is necessary to understand the atmospheric chemistry of SO2 in order to understand the role of biogenic sulfur in the atmosphere. In this paper, we will discuss the current understanding of the homogeneous gas phase oxidation of SO2 by hydroxyl radicals (HO) in the atmosphere. This 1

1

1

0

0097-6156/89/0393-0437$06.00/0 1989 American Chemical Society

Saltzman and Cooper; Biogenic Sulfur in the Environment ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

438

BIOGENIC SULFUR IN THE ENVIRONMENT

reaction ultimately leads to the formation of sulfuric acid (H2SO4) and is of major importance to acidic deposition processes. The reaction of HO with SO2 is believed to proceed by the addition of the HO radical to S 0 (2) 2

HO + S 0 (+ M) - > H O S 0 (+ M) 2

(1)

2

Until recently, our knowledge of the fate of the H O S 0 radical under atmospheric conditions was limited. To avoid the uncertainties in the chemistry of the HOSO2 radical, early atmospheric chemical models for acid deposition (2) have treated the process as 2

HO + S 0 -> -> -> H S 0 Downloaded by UNIV OF IOWA on September 2, 2016 | http://pubs.acs.org Publication Date: April 27, 1989 | doi: 10.1021/bk-1989-0393.ch026

2

2

(2)

4

Chemical modeling using this simplified mechanism (2) for H S 0 formation has shown that there are nonlineanties in the changes m H0SO4 concentrations calculated as the S 0 concentration is changed (3-5). When the SO2 concentration is reduced by 30%, the H7SO4 concentration is only reduced by 14%, not the 30% reduction expected if there were a linear relationship (1). This nonlinearity is a major complication in relating S 0 emissions to H 2 S O 4 deposition. Reaction (2) leads to the loss of one H O radical for each S 0 molecule oxidized, and the formation of one molecule of H S 0 . As the SO? concentration decreases, this mechanism suggests that the H O radical concentration must increase, thus increasing the efficiency of the SOo to H S 0 conversion. Although the amount of H S04produced increases, it does not do so in direct proportion to the S 0 increase. Thus, this mechanism is responsible for the nonlinearity in the relationship between S 0 concentration changes and H S 0 concentration chances. In 1983, Stockwell and Calvert (&) presented data which suggested that an alternate mechanism might be more appropriate for the description of the H O - S O ? reaction i n air. In their experiments, they photolyzed H O N O / N O x / C O mixtures in the presence and absence of SO?. Data were collected on the C 0 production rate from the photooxidation of CO. This is directly related to the H O radical concentration in the reaction system. The data suggested that there was no significant change in the H O concentration upon the addition of S 0 to the reaction chamber. They interpreted their data with the following mechanism: 2

4

2

2

2

2

4

2

4

2

2

2

2

4

2

2

HO + S 0 (+ M) - > H O S 0 (+ M)

(1)

H O S 0 + 02 - > H 0 + S O 3 2

(3)

S0

4

(4)

2

2

2

+ H 0->H S0

3

2

2

In their system, sufficient nitric oxide (NO) was present to react with the H 0 radical, reforming the HO radical. H0

2

+ NO - > H O + N 0

2

2

(5)

These data and the resulting mechanism are consistent with earlier observations of smog chamber processes conducted in the presence and absence of added S 0 , as discussed in references (2 6). In the atmosphere, reaction (4) is expected to be the primary fate of SO3 aue to the presence of large quantities of water and the presumed lack of other important reactions for SO*. Under atmospheric conditions, this mechanism suggests that H O radicals are effectively converted to H 0 radicals, if reaction (3) is sufficiently fast and there 2

2


26. ANDERSON ETAL.

Oxidation ofSO by OH

439

2

are no other significant loss processes for the H O S 0 radical. Thus the net reaction under atmospheric conditions becomes 2

H O + S 0 (+ O * H 0 ) - > H S 0 + H 0 2

2

2

4

(6)

2

Meagher et al. (2) performed similar experiments in propene/butane/NOx/ H 0 mixtures with and without added S 0 . These experimental results were compared with model results from the use of reaction (2) or the use of reaction (6) to describe the H O + S 0 reaction. The experimental results compared reasonably well with the model results from the use of reaction (6), ana very poorly with thosefromthe use of reaction (2). Three studies have reported measurements of the rate constant for reaction (3), H O S 0 + Oj. These three studies approached the problem differently, and will be discussed individually. In the earliest of these studies, Marritan (g) used a flash photolysis/resonance fluorescence system for the study of the H O radical reaction with S 0 . The H O radical decay was followed as a function of time with S 0 present and with and without NO and O? added. It was observed that the H O radical decay rate was smaller when N O and 0 were added, suggesting that H O radicals were being regenerated when both NO and 0 were present. These experiments demonstrated the necessity for the presence of both NO and O^ for this H O radical regeneration. These experimental results were fit with a simple chemical mechanism and the rate constant for reaction (3), H O S 0 + 0 , was determined to be (4 ± 2) x IO" cm /s. It was also found that the H O S 0 radical reacted with NO with a rate constant of about 2.5 x IO" cm /s. The interpretation of the experimental data was complicated by the fact that all of the reactants were present throughout the reaction, hence many reactions occur simultaneously. In addition, the H O radical concentration in this complex reaction system was monitored as a function of reaction time for only about 10 ms. Schmidt et al. (2) have reported very similar experimental results, but have not reported a rate constant for the reaction of H O S 0 with o . The second study of the kinetics of the H O S 0 reaction with 0 was reported by Martin et al. QQ). They used discharge flow techniques to produce the H O radicals and E P R detection of the H O radicals. Some of their experiments were carried out in a manner analogous to Margitan's. Namely, S u , 0 and NO were added to the flow system through the same movable injection port. Under these conditions all reactions were occurring simultaneously. They also performed experiments in which they took advantage of one of the characteristics of flow systems, the ability to add reactants at different reaction times. In these experiments, they introduced S 0 and O? through the same movable port and added NO through a fixed port upstream of the EPR detector. Typical reaction times used in tneir system were about 20 ms. Chemical modeling was used to estimate the rate constants for various reactions occurring in the system. The H O S 0 + O? reaction (3) was reported to have a rate constant of (3.5 ± 1) x IO" cm /s, the H O S 0 + NO reaction was ^ 5 x IO* cm /s. In the studies discussed above, HO radicals were detected. H O radicals are neither a reactant nor a product of the H O S 0 + 0 reaction (3). Hence, these studies of the kinetics of reaction (3) are indirect. In the work of Gleason et al. (H), the reactant H O S 0 and the product SO3 were monitored. This work at first glance appears to be a direct study of the kinetics of reaction (3), until one realizes that these species were monitored by chemical ionization mass spectrometry. H O S 0 was converted to SO3- and S O 3 was converted to (C1-S03)- prior to mass spectrometric detection. The detection of H O S 0 and S O 3 are based upon a complete understanding of the chemical ionization 2

2

2


2

2

2

2

2

13

3

2

2

2

12

3

2

2

2

2

2

2

2

2

13

3

2

13

3

2

2

2

2

2


440


processes, much in the same way that the previous studies used the understanding of the H 0 + NO reaction to allow one to measure H O and infer a knowledge of the HO? concentration. This chemical ionization technique should not be considered a direct spectroscopic measurement of HOSO2 SO3. This study provides the first evidence that SO3 is formed as a result of reaction (3). From this work, the rate constant for the H O S 0 + 0 reaction is reported as (4.37 ± 0.66) x IO" cm /s. They did not measure the rate constant for the HOSO? + NO reaction. In their paper, they also report the results of a study by Bandow and Howard of the H O S 0 + O? reaction. In that work, direct spectroscopic measurement of the other product H 0 was reported using laser magnetic resonance techniques. The rate constant for reaction (3) was reported to be (4.9 ± 2.9) x IO" cm /s. Several modeling studies have been performed to evaluate the difference between using reaction (6) and reaction (2) to represent the oxidation of S 0 by H O in atmospheric model calculations. Samson (12) has performed calculations which show a more nearly linear relationship between the change in S 0 emissions and the change in H S04 concentrations when using reaction (6) than is found by the use of reaction (2). This is the case in model calculations that at least crudely treat the aqueous phase oxidation of S 0 , as well as the gas phase oxidation. Seigneur et al. (12) have shown that by using reaction (6) to describe the oxidation of S 0 by HO in an atmospheric chemical model produces a very nearly linear relationship between sulfate concentrations calculated and S 0 precursor concentrations under clear-sky conditions. Under cloudy conditions, when the aqueous phase processes become more important, the response to the S 0 precursor concentration changes is no longer linear. Stockwell et al. (14) have shown that by using reaction (6) to describe the H O radical oxidation of S 0 to H S 0 in the more complete gas phase reaction mechanism of the National Center for Atmospheric Research's Regional Acid Deposition Model (RADM) good linearity is observed between the sulfate produced and the S 0 precursor concentrations under both rural and urban atmospheric conditions. By using reaction (6) to describe the gas phase oxidation of S 0 to sulfate, a variety of different models seem to show good linearity between the S 0 precursor concentrations and the sulfate produced. It is therefore expected that much of the nonlinearity that may actually occur in the atmosphere between S 0 emissions and sulfate concentrations arises from the aqueous phase or cloud processes for the conversion of S 0 to sulfate. The justification for using reaction (6) to describe the gas phase oxidation of S 0 to sulfate, is based on the belief that the H O S 0 radical reacts quickly with 0 to form HO? and SO3, and that there is no other significant loss process for the H O S 0 radical. In the current study, a discharge flow-resonance fluorescence system was used to produce and monitor the concentration of HO radicals. The H O radical signal decreases as a function of reaction time upon the addition of S 0 to the system, reaction (1). The kinetics of this reaction have been studied extensively and have not been reinvestigated in this work. Rather, the subsequent reactions have been investigated in the presence of added NO, or O?, or NO and 0 to determine the importance of the proposed mechanism (1), (3) and (5). The rate constants for the H O S 0 reactions with 0 and NO are reported. 2

o r

2

13

2

3

2

2


13

3

2

2

2

2

2

2

2

2

2

4

2

2

2

2

2

2

2

2

2

2

2

2

2

Experimental

A discharge flow-resonance fluorescence system has been constructed for use in these stumes. A mixture of H in He was passed through a microwave discharge to produce H atoms. Very low concentrations of H atoms can be generated by passing pure He through the discharge. In all of the experiments described in 2

M

H


26. ANDERSON ETAL.

Oxidation of SO by OH

441

2

this work, He was used as the diluent and the reactions were studied at a pressure of about 3 torr. A short distance downstream of the H atom source, a mixture of NO? in He was added to the flow. The H atoms react quantitatively with NO? to produce HO radicals. Further downstream the reactant gases SO?, 0 and NO were added to the reaction system. Each of these gases were added to the system through a series of fixed ports, rather than a movable injector. The principal advantage of this technique is that as the position at which reactants are added to the system is changed, one is not changing the surface area to which the radicals are exposed. In flow systems using movable injectors a correction must be applied to the data to account for radical losses on the changing wall surface area. This type of correction is unnecessary for flow systems using fixed ports. In the detection zone of the flow system, the H O resonance fluorescence signal is directly proportional to tne H O radical concentration.


2

Resets and Discussion Experimental Results. Most of the experiments that have been performed in the current studies were carried out with the sequential addition of reactants to the reaction system. In the experiments to be discussed in this work, the SO? is introduced into the flow system about 115 ms upstream of the H O detection region. The 0 was added to the reaction system about 95 ms upstream of the HO detection system. Finally, NO was added to the reaction system at about 15 ms upstream of the detector. Alternatively for studying the H O S 0 + NO reaction, the NO was added to the flow system about 135 ms upstream of the H O detection system, and upstream of the S 0 inlet. By carrying out the reactions in a sequential manner, the chemistry (or at least tne more important reactions) occurring in different regions of the flow system were simplified. Most of the experiments were performed by adding about 0.5 - 2 x 10 molecule/cm of SO? to the flow system. In this concentration range, 80% or more of the H O radicals were removed from the reaction system by reaction with SO? prior to the H O detection region. Experiments were performed in which the amount of 0 added to the reaction system varied from 0 - 2.0 x 10 molecule/cm . N O was added to the flow system at relatively low concentrations, about 1.0 x 1 0 molecule/cm . This concentration was sufficiently high to insure virtually complete conversion of H 0 to H O in the approximately 15 ms reaction time. The NO concentration was sufficiently low and the reaction time was sufficiently short, that there was little effect on the HO radical concentration due to its reaction with NO to form HONO. Figure 1 shows a plot of a typical set of data. These data were collected with a relatively high initial HO radical concentration of about 5 x 10 /cm at the SO? inlet port. The S 0 concentration was 1.1 x 10 molecule/cm and the initial r J 0 concentration was 3.3 x 10 molecule/cm . Molecular oxygen was added at concentrations between 0 and 2.2 x 10 * molecule/cm and NO was added 15 ms prior to the detection system at a concentration of 1.2 x 10 molecule/cm^. At this high concentration of SO? and the relatively long reaction time employed, the H O radical concentration remaining at the detector is about 1.1 x 10 /cm , in the absence of added 0 . As increasing quantities of 0 are added to the reaction system, the H O radical concentration increases to about 3 x 10 /cm when the 0 concentration reaches about 1 x 1 0 molecule/cm . These data slightly underestimate the H O radical concentrations in the presence of 0 , since the H O radical fluorescence intensity has not been corrected for the effect of fluorescence quenching by 0 . The 0 quenching effect could amount to as much as 10% at the highest O? concentrations used. Each data point shown in the figures is the weighted 2

2

2

15

3

15

2

3

14

3

2

11

15

3

3

2

12

3

2

1

3

14

10

3

2

2

10

3

2

15

3

2

2

2


442



5

0 I 0

i 5

I

I

I

10

15

20

I 25

3

0z (x 10** moleculc/cm ) Figure 1. Experimental HO radical concentration data and model results as a function of added 0 concentration. Initial H O radical concentration - 5 x 10 /cm at the SO? port, N 0 concentration - 3.3 x 10 molecule/cm S 0 added 115 ms from HO detection zone at concentration -1.1 x 10 molecule/cm NO added 15 ms from H O detection zone at concentration - 1.2 x 1 0 molecule/cm and 0 2 added 95 ms from HO detection zone at concentrations 0 - 2.2 x 10 molecule/cm . The curves show the effect of varying the H O S 0 + O? rate constant upper - 8.0 x IO cm /s, middle - 4.0 x IO" cm /s and lower ZOxlO-ttcmVs. 11

2

3

12

3

2

2

15

3

14

3

15

3

2

-13

3

13

3


26. ANDERSON ETAL.

Oxidation of S0 by OH

443

2

average of the between two and twenty individual data points collected in a series of similar experiments. Each of the individual data points is the average of ten measurements of the H O radical fluorescence intensity, and hence the H O radical concentration. The individual data points are weighted by the reciprocal of the standard deviations in these data points. The error bars shown in the figures were determined from the standard deviations in the measured H O fluorescence signals and an assessment of the absolute accuracy of the measurement of the H O concentration. From this data, it is clear that the H O radical concentration increases as 0 is added to the flow system. Similar experiments were performed in the absence of added NO. These experiments show only a very small increase of the H O signal due to the addition of increasing quantities of 0 to the reaction system. This small effect remains because there is a small quantity of NO present in our reaction system, from the initial production of the HO radical. Figure 2 shows the results for a similar set of experiments in which the initial H O radical concentration was about 4.S x 10 / c m at the S 0 port, the initial N 0 concentration was 3.3 x 10 molecule/cm , the S 0 concentration was 1.0 x 1 0 molecule/cm , the N O concentration was 1.1 x 1 0 molecule/cm and the 0 2 concentration was between 0 and 2.0 x 1 0 molecule/cm . This time the NO was added about 135 ms upstream of the HO detection zone. Under these conditions the observed signal will be affected by both the reaction of the H O radical with NO and by the reaction of H O S 0 with NO. The H O radical concentration at the detector is only about 0.5 x 10 /cm in the absence of added O?, rising to about 1.5 x 10 /cm when the 0 concentration exceeds about 1 x 10 molecule/cm Figure 3 shows a plot of experimental data for a set of experiments in which the initial H O radical concentration was about 3 x 10 / c m at the S 0 port, the initial N 0 concentration was 3.3 x 1 0 molecule/cm , the S 0 concentration was 1.0 x 10 molecule/cm , the O^ concentration was 1.1 x 10 molecule/cm and the NO concentration was varied between 0 and 3.2 x 10 molecule/cm . For this set of experiments the NO was added about 135 ms upstream of the H O detection zone. In these experiments, the H O radical concentration was observed to increase with the smallest quantities of added NO. As the N O concentration was increased further, the H O radical concentration was observed to decrease. Several other sets of experiments of the types shown in the figures were performed using different initial concentrations of the H O radical and SO?. These figures show an example of the types of experimental results and analysis that have been performed. The conclusions presented are based on the analysis of the results over a broader set of experiments. 2


2

11

3

2

12

3

2

2

15

3

14

3

15

3

2

10

3

10

15

3

3

2

11

3

2

12

3

2

15

2 15

3

3

14

3

Model Results. In order to interpret the results of our experiments, we have undertaken an extensive chemical modeling effort to investigate the chemistry occurring in this reaction system. The reactions used in this model are listed in Table I. This is a more complete mechanism than that used by either Margitan (8) or Martin et al. (1Q) in their studies of the kinetics of this reaction system. Gleason et al. (H) reported no modeling of their studies. The rate constants for many of the reactions listed in Table I are known (15). The wall loss rate constant for H O radicals used initially in the model was based upon measurements made in our reaction system. The H 0 and HOSO? wall loss rate constants were determined in our modeling studies. The data of Martin et al. (1Q) were used to calculate the rate constant for the reaction of HO with S 0 which was appropriate for our experimental conditions. Their rate constant results allowed us to calculate an accurate effective bimolecular rate constant for this reaction in the presence of large quantities of S0 . The initial choice of 2

2

2



444


E \U a 3 U a o

E

O X

02

(x 10

molecule/cm )

Figure 2. Experimental HO radical concentration data and model results as a function of added 0 concentration. Initial H O radical concentration - 4.5 x 10 /cm at the S 0 port, N 0 concentration - 3.3 x 10 molecule/cm , SO? added 115 ms from H O detection zone at concentration - 1.0 x 10 * molecule/cm , NO added 135 ms from H O detection zone at concentration 1.1 x 10 molecule/cm and O? added 95 ms from H O detection zone at concentrations 0 - 2.1 x 10 molecule/cm . The curves show the effect of varying the H O S 0 + NO rate constant upper - 0.5 x IO cm /s, middle -1.0 x IO* cm /s and lower - 2.0 x IO" cm /s. 2

11

3

12

2

3

2

1

3

14

3

15

3

-12

3

2

12

3

12

3


26. ANDERSON ETAL.

Oxidation of S0 by OH

445

2

5


4

0

I

1

1

0

.8

1.6 NO

I—I

1

2.4

II

3.2

4

(x 1 0 * * m o l e c u l e / c m ) 3

Figure 3. Experimental HO radical concentration data and model results as a function of added NO concentration. Initial HO radical concentration - 3 x 10 /cm at the SO? port, N 0 concentration - 3.3 x 10 molecule/cm , S 0 added 115 ms from H O detection zone at concentration -1.0 x 10 molecule/cm ,0? added 95 ms from H O detection zone at concentration - 1.1 x 10 * molecule/cm and NO added 135 ms from HO detection zone at concentrations 0 - 3.2 x 10 molecule/cm . The curves show the effect of varying the HOSO? + NO rate constant upper - 0.5 x IO* cm /s, middle -1.0 x IO-* cm /s and lower-2.0 x l O - ^ c m ^ . 11

3

12

3

2

2

15

3

1

3

14

3

12

3

2

3


3


446

Table I. Chemical Mechanism Used to Describe Chemistry Occurring in the Discharge Flow System Rate Constant cm molecule" s"

Reaction


3

1

1) H + N 0 - > H O + NO 2) H O + NO (+ M) - > HONO (+ M) 3) H O + N 0 ( + M) - > HONO2(+ M) 4) HO + H O - > H O + O 5) O + H O - > H + 0 6) 0 + N 0 2 - > 0 + NO 7) H O + wall - > products 8) H O + S 0 + M - > HOSC^ + M 9) H O S 0 + 0 - > H 0 + SO3 10) H O S 0 + HO ~> products 11) H O S 0 + NO ~> products 12) H 0 + N O - > H O + N 0 13) H02 + N 0 ~ > H 0 N 0 14) O + H 0 - > H O + 0 15) H O + H 0 2 - > H 0 + 0 16) H C ^ + H 0 ~> H 0 + 0 17) H 0 + wall - > products

1.13 x l O " 6.9 x l O " 2.5 x l O " 1.9 xlO* 3.3 x l O " 9.3 x l O "

18) H O S 0 + wall - > products

See text

2

2

a

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

Reference

1

10

14

13

12

11

12

15 15 15 15 15 15

1

18 s1.2 x l O " See text 5 x IO" See text 8.3 x l O " 2.0 x l O 5.9 x l O " 7.0 x l O 1.7 x l O See text

14

10 10

11

12

14

11

11

12


15 15 15 15 15

26. ANDERSON ETAL.

Oxidation ofSO by OH

447

2

the rate constants for the HOSO2 reactions with H O radicals and N O were taken from Martin et al. (1Q). Comparison of Model and Experimental Results. Modeling a wide variety of different experimental conditions while varying the rate constant for the H O + HOSO2 reaction, demonstrated that the results were not significantly affected by the choice of this rate constant up to a value of 1 x I O cmP/s. The experimental results shown in Figure 1 were obtained with NO added to the reaction system for the final IS ms prior to HO detection. Model results for this type of experiment were found to oe largely unaffected by the choice of the rate constantforthe HOSO2 + NO reaction. Figure 1 shows a comparison of the model results with the experimental results. The three curves shown in the plot correspond to three different values of the rate constant for the HOSO? + 0 reaction; upper - 8 x IO , middle - 4 x IO* , and lower - 2 x K h cm /s. Similar comparisons between model and experimental results have been made for a wide variety of other experimental conditions. Based upon such comparisons, we have concluded that a rate constant of (4 x IO* cm /s gives the best match between the experimental and model results, in both an absolute sense and based upon the shape of the 0 titration results. Since there is greater uncertainty in the absolute concentrations of H O radicals than there is in the trend of the H O concentrations with increasing 0 , the comparison of the shapes of the experimental and model O2 titration profiles may provide a reliable basis for comparison. From modeling a variety of different experiments with and without SO2 added to the system, it was found that the H O radical wall loss must increase when SO2 is added to the reaction system. Experimentally we found that an H O wall loss rate constant of 10/s worked quite well in the models of experiments performed in the absence of SO2. When SO2 was added to the reaction system, it was necessary to increase the HO wall loss rate constant used in the model to (18 ± l)/s to be able to adequately describe the experimental results. The comparison between the experimental and model results shown in Figure 1 are based on the use of a H O S u wall loss rate constant of 30/s. Our modeling studies have shown that changing the H O S 0 wall loss rate constant by ± 10/s has little effect on the overall comparisons between the experimental and model results. The HO? wall loss rate constant used when SO? was present in the reaction system was 20/s. There was no effect on the modeling results of varying this wall loss rate constant by ± 10/s. Figure 2 again shows a comparison between experimental and model results. In this experiment, the NO was added to the reaction system about 135 ms upstream of the H O radical detection system. In this case the effect of varying the rate constant for the HOSO2 + N u reaction is being shown. Three different rate constants were used for the HOSO2 + NO reaction; upper - 0.5 x IO- , middle -1.0 x IO* , and lower - 2.0 x IO" cm /s. From several different comparisons between the experimental and model results, we have concluded that the rate constant for the H O S 0 + NO reaction is best described as (1 tj) x IO* cm /s. The data shown in Figures 1 and 2 are for essentially the same concentrations of reactants, only the reaction time for the N O in the system is different. The data shown in Figure 2 is more representative of the results that would be obtained if all of the reactants were present throughout the entire reaction period. The model results for these experiments are significantly affected by the choice of rate constants for both the H O S 0 reactions with 0 and NO. By performing experiments such as those shown in Figure 1, one is able to experimentally study a kinetic system which depends -10

-13

2


13

1 3

13

3

3

2

2

2

2

12

12

12

3

2

12

3

2

2

American Chemical Society Library 1155 Biogenic 18th St., M.W. Saltzman and Cooper; Sulfur in the Environment ACS Symposium Series;Washington, American Chemical D.C.Society: 20036Washington, DC, 1989.

448


significantly on only one unknown rate constant. This demonstrates one of the advantages of flow studies of complex kinetics. Figure 3 shows the comparison between the experimental and model results for a variation in the rate constant for the HOSO2 + NO reaction between 0.5 and 2.0 x 10" cm /s. These experiments were performed with 1.1 x 10 molecule/cm of 0 and a variable concentration or NO added to the reaction system 135 ms upstream of the H O detection system. The initial increase in the H O concentration upon addition of NO is due to the introduction of sufficient NO to drive the conversion of H 0 to H O to completion. The further decay in the H O concentration with added NO is due to the combined effects of the HO reaction with NO and the HOSOo reaction with NO. Again this data suggests that the rate constant for the HOSOo + NO reaction is adequately described as (11\) x IO* cm /s. This value of the rate constant is smaller than the 2.5 x IO* cm /s reported by Margitan (g), and larger than the < 5 x IO' cm /s reported by Martin et al. (10). Our value for the HOSO* + NO rate constant is only about 2.5 times larger than that for the HOS0 + O2 rate constant, hence the reaction of HOS0 with NO will be of no importance in the atmosphere. The only other constituent in the atmosphere that is present in sufficient concentration to possibly compete with Ch is a possible reaction with HOS0 is water vapor (3.9 x 10 molecule/cm at 5b % R.H.). If the rate constant for an HOSOb + H 0 reaction were about 5 x I O cm^s, the reaction with water would be about equally important as the reaction of HOS02 with 0 . We are currently performing experiments which will allow us to evaluate the possible effects of H 0 on the HOS0 + 0 reaction. If the hydration of the HOS0 radical is as important as Friend et al. (1&) suggest, the SO? oxidation mechanism might be altered by the presence of H D . None of the previous kinetics studies have been performed in the presence of H 0 , hence there is no basis for evaluating the potential importance of the HOS0 »OH complex under atmospheric conditions. In summary, this work suggests that the rate constant for the HOS0 + 0 reaction is (4 t$) x IO" cm /sand that the rate constant for the HOS0 + NO reaction is (1 f}) x IO" cm /s. This rate constant for the HOS0 + NO reaction is too slow for this reaction to ever be of any significance in the atmosphere, when compared to the HOS0 + 0 reaction. 12

3

15

3

2

2

12


12

3

3

13

3

2

2

2

17

3

-12

2

2

2

2

2

2

2

2

2

2

2

13

2

3

2

12

3

2

2

2

Acknowledgments The authors would like to thank Juan L. Bonilla for his assistance with the data acquisition system and software development for these experiments. This research is based upon work supported by the National Science Foundation under Grants ATM-8405394 and ATM-8521192.

Literature cited 1. Toon, O. B.; Kasting, J. F.; Turco, R. P.;Liu,M.S. J. Geophys. Res. 1987, 92D, 943-63. 2. Calvert, J. G.; Stockwell. W. R. In Acid Precipitation - Vol. 3: SO , NO and NO Oxidation Mechanisms: Atmospheric Considerations; Ann Arbor Science Publishers Ann Arbor, MI, 1983, pp 1-62. 3. Rodhe, H.; Crutzen, P.; Vanderpol, A. Tellus 1981,33,132-41. 2

2


26. ANDERSON ETAL.

Oxidation of S0 by OH 2

449


4. Acid Deposition: Atmospheric Processes in Eastern North America. National Academy of Sciences, Washington, D.C., 1983. 5. Samson, P. J.; Small, M . J. In Acid Precipitation - Vol. 9: Modeling of Total Acid Precipitation Impacts; Ann Arbor Science Publishers Ann Arbor, MI, 1984, pp 1-24. 6. Stockwell, W. R.; Calvert, J. G. Atmos. Environ. 1983, 17, 2231-5. 7. Meagher, J. F.; Olszyna, K. J.; Luria, M . Atmos. Environ. 1984, 18, 2095104. 8. Margitan, J. J. J. Phys. Chem. 1984, 88, 3314-8. 9. Schmidt, V.; Zhu, G. Y.; Becker, K. H.; Fink, E. H . Ber. Bunsen-Ges. Phys. Chem. 1985, 89, 321-2. 10. Martin, D.; Jourdain, J. L.; Le Bras, G. J. Phys. Chem. 1986,90,4143-7. 11. Gleason, J. F.; Sinha, A.; Howard, C. J. J. Phys. Chem. 1987,91,719-24. 12. Samson, P. J., as quoted in Acid Deposition: Atmospheric Processes in Eastern North America, National Academy of Sciences, Washington, D.C., 1983 and Regional Acid Deposition: Models and Physical Processes, NCAR/TN-214+STR, Boulder, CO, 1983. 13. Seigneur, C.; Saxena, P.; Roth, P. M . Science 1984, 225, 1028-9. 14. Stockwell, W. R.; Milford, J. B.; McRae, G. J.; Middleton, P.; Chang, J. S., accepted for publication in Atmos. Env. 1988. 15. DeMore, W. B.; Margitan, J. J.; Molina, M . J.; Watson, R. T.; Golden, D. M . ; Hampson, R. F.; Kurylo, M . J.; Howard, C. J.; Ravishankara, A . R. Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling. Evaluation Number 7; J P L Report 85-37, Jet Propulsion Laboratory Pasadena, CA, 1985. 16. Friend, J. P.; Barnes R. A.; Vasta, R. M . J. Phys. Chem. 1980, 84, 2423-36. RECEIVED August 5, 1988


Mechanism of Atmospheric Oxidation of Sulfur Dioxide by Hydroxyl

Recommend Documents